MRI-Based Tumour Targeting Enhancement with Magnetotactic Bacterial Carriers

(1)

UNIVERSITÉ DE MONTRÉAL

MRI-BASED TUMOUR TARGETING ENHANCEMENT WITH

MAGNETOTACTIC BACTERIAL CARRIERS

OUAJDI FELFOUL

INSTITUT DU GÉNIE BIOMÉDICAL ÉCOLE POLYTECHNIQUE DE MONTRÉAL

THÈSE PRÉSENTÉE EN VUE DE L’OBTENTION DU DIPLÔME DE PHILOSOPHAE DOCTOR (Ph. D.)

(GÉNIE BIOMÉDICAL) JUILLET 2011

(2)

UNIVERSITÉ DE MONTRÉAL

ÉCOLE POLYTECHNIQUE DE MONTRÉAL

Cette thèse intitulée:

MRI-BASED TUMOUR TARGETING ENHANCEMENT WITH

MAGNETOTACTIC BACTERIAL CARRIERS

présentée par: FELFOUL Ouajdi

en vue de l’obtention du diplôme de: Philosophae Doctor a été dûment acceptée par le jury d’examen constitué de:

Mme CHERIET Farida, Ph.D., présidente

M. MARTEL Sylvain, Ph.D., membre et directeur de recherche M. SYLVESTRE Michel, Ph.D., membre

(3)

ACKNOWLEDGMENT

The completion of this work has not been possible without the scientific, moral and human support that I received from several persons to whom I am very grateful. It remains certain that my supervisor, Professor Sylvain Martel, whose strategic vision and the human approach were able to help me to succeed. Professor Sylvain Martel, thank you for having opened the doors of your laboratory where I learned to do things differently and to try what others dare not explore.

I thank the jury: Prof. Cheriet, Prof. Ferreira, Prof. Sylvestre, and Prof. Henry for having agreed to evaluate this thesis. You honor me by evaluating this work. Please accept my respectful thanks.

A very special thought Dr. Mahmood Mohammadi, whose expertise was essential to the accomplishment of this work. With your magic fingers, we have undertaken great scientific discoveries.

I thank Dr. Louis Gaboury (UdeM), for the many tips in the analysis of histological sections, Micheline Fortin (UdeM), Julie Hinsinger (UdeM) and the team of histology (UdeM), for the preparation of tissue and interest in the project.

I thank Hakim Slimani, from the Institute Armand Frappier, and Hélène Ste-Croix, from MethylGene. Your support has enabled the implementation of in vivo tests that were a turning point in my work.

A very special tribute to the members of the laboratory Nanorobotics. It was a great pleasure to have attended over the years. Friendships were forged which I hope will last for years and years

(4)

to come. Neil Kaou, whose presence in the laboratory provides support, joy and organization. For Nasr, Dominic, Nisryn and Charles, for their friendship and continuous support.

I thank my family for their support during these long years of study and without which I would not be here today. To my family: mami and papi, Lahouma, Tata, Mahdouch, Halloula, Laysoun, Nanny, Nanna, Kimou, Mohsen Felfoul Junior, and Ammounatou you have been a continuous source of inspiration that has done much to help me get through requirements and the ups and downs of this thesis. Without you, I would probably have been discouraged by the difficulties associated with these academic activities.

The consecration of my studies and the culmination of many years dedicated to research coincided with the knowledge of the woman of my life, Besma that gived me a strong push at the end of the thesis.

This project is supported in part by a Canada Research Chair (CRC) in Micro/Nanosystem Development, Fabrication and Validation, the Canada Foundation for Innovation (CFI), the National Sciences and Engineering Research Council of Canada (NSERC), and the Fonds Québécois de Recherche sur la Nature et les Technologies (FQRNT). This work was also supported by US grant Number R21EB007506 from the National Institute Of Biomedical Imaging And Bioengineering. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Imaging And Bioengineering or the National Institutes of Health.

(5)

ABSTRACT

Magnetotactic Bacteria (MTB) are being explored as potential drug transporters to solid tumours. The MTB’s active motility combined with magnetotaxism (their ability to swim following the direction of a magnetic field) offer new and potentially more accurate solutions in delivering drugs to tumours. In fact, the flagella bundles of the MC-1 bacteria (with an overall ideal cell diameter of approximately 50% the diameter of the tiniest human blood vessels) provide 4.0 to 4.7pN of thrust force for propulsion (roughly 10 times the value of many other well-known flagellated bacteria). Since there are no existing methods or technologies capable of inducing an equivalent force on a carrier of appropriate size for traveling inside a tumour’s microvasculature, live microorganisms are considered as a viable option. Many of the parameters in a tumour microenvironment, such as malformed angiogenesis capillaries, heterogeneous blood flow, and high interstitial pressure, hinder the delivery of blood-borne drugs to the affected area. Active motility might prove to be helpful in bypassing these limitations and may facilitate the uniform distribution of the drug in the targeted area.

An MTB navigation technique that allows targeting without prior knowledge of the exact architecture of the vessels network has been developed. This navigation technique exploits both the ability of the MTB to swim following an imposed magnetic field and their random, continuous motion at low magnetic fields. Firstly, a focused magnetic field on the target sets the overall direction of the bacteria. Then, as the bacteria approach the targeted zone, the intensity of the magnetic field is decreased, which allows better bacteria repartition by exploiting their free motion. An additional approach that enhances MTB targeting relies on modulating the magnetic field direction in time, while keeping the magnetic field lines pointed toward the target. Navigation experiments in complex micro-channel networks highlight this process, where the successful targeting of bacteria is demonstrated when an appropriate magnetic field algorithm is applied, especially when it takes into account the nature of the channel network. Tridimensional control and navigation of MTB is also possible with the same technique through proper powering of the magnetic coils. In fact, by controlling their magnetic environment, it is possible to form a swarm of MTB, control its size and position within a given volume using a computer program.

(6)

MTB respond to magnetotaxis-based directional control since each cell contains internal structures known as magnetosomes. In the case of the MC-1 cells, these structures are composed of magnetite (Fe3O4) nanoparticles arranged in a chain that acts similarly to a magnetic nano-compass. The magnetosomes embedded in each MTB can be used to track the displacement of these bacteria using a Magnetic Resonance Imaging (MRI) system. These magnetosomes disturb the local magnetic field and affect the T1 and T2-relaxation times during MRI. Magnetic Resonance (MR) T1- weighted and T2-weighted images as well as T2-relaxivity of MTB are studied in order to validate the possibility of monitoring MTB drug delivery operations using a clinical MR scanner. It is found that MC-1 MTB affect the T2-relaxation much more than the T1-relaxation rate and can be thought of as a negative contrast agent. The signal decay in the T2-weighted images is found to change proportionally to the bacterial concentration. A minimum concentration of 2.2×107_{cells/mL can be detected with a standard 1.5Tesla clinical MR system} using a T2-weigthed image. Furthermore, when the influence of the magnetosome’s chain, the motility, and the bacterial cell of MC-1 MTB on the MRI contrast were studied, it was found that nanoparticles synthesized by MC-1 MTB were the predominant source of contrast in MRI. In vivo studies, invastigating the ability of MC-1 MTB to reach the tumour, revealed the presence of these bacteria in the necrotic zone of solid tumours. The application of a magnetic field focused on the target, using a special magnetic system that was designed and built especially for these in vivo experiments, helped the accumulation of the bacteria in the tumour. In this paper we report on average, twice the number of bacteria found in the tumour when a magnetic field was applied using the 3D magnetic coils system over passive transport with the blood flow in the absence of any magnetic field. A second experience based on the comparison of tumours implanted in the same animal reported 10 times more bacteria in the targeted over the non-targeted tumour, which may suggest that the applied magnetic field may also be used to avoid a specific zone in the body.

(7)

RÉSUMÉ

Le cancer constitue la première cause de mortalité au Québec, avec 20,000 décès estimés par année. Parmi tous les patients atteints du cancer, une grande proportion pourrait profiter de l’avancement technologique en ce qui concerne le transport de médicaments. En effet, l’un des meilleurs moyens d’augmenter l’efficacité d’un médicament contre le cancer, tout en réduisant sa toxicité sur les cellules saines, est de le diriger vers la tumeur et de le maintenir à cet endroit jusqu’à ce qu’un effet thérapeutique se produise. Le transport ciblé de médicaments vers la tumeur peut considérablement améliorer l’efficacité thérapeutique, surtout si le transporteur est capable d’atteindre les zones nécrotiques et se répartir uniformément dans la zone à traiter. Les bactéries, de par leur motilité, sont d’excellents candidats pour une telle application, surtout qu’elles peuvent aussi être facilement fonctionnalisées. Ainsi, la recherche sur le traitement du cancer utilisant des bactéries s’est imposée comme une approche prometteuse surtout qu’elle pallie à une limitation majeure de la chimiothérapie et de la radiothérapie en permettant le traitement des zones anaérobies.

Alors que des laboratoires à travers le monde tentent de fabriquer des systèmes miniatures en se basant sur le modèle bactérien, nous avons opté pour l’utilisation des bactéries qui existent dans la nature. Notre stratégie a été de trouver un système biologique ayant les caractéristiques essentielles (e.x. diamètre total de moins de deux micromètres, force de poussée de plus de 4 pN, etc.) et de concentrer nos efforts à identifier une interface et une méthode permettant son contrôle pour des fins de ciblages thérapeutiques dans les lésions tumorales. Nous avons identifié les bactéries magnétotactiques de type MC-1 comme le meilleur transporteur potentiel de médicaments pour le ciblage du cancer.

Les MC-1 sont à la fois dirigeables par champs magnétiques et anaérobies, ce qui leur donne un grand avantage par rapport aux bactéries traditionnellement utilisées pour le ciblage du cancer. Le ciblage du cancer avec des bactéries exploite le plus souvent l’affinité des bactéries anaérobies à la région nécrotique faible en oxygène de la tumeur. Certes, ce ciblage manque de spécificité et un des problèmes le plus reconnu est la nécessité d’injecter une forte dose de bactéries pour assurer une croissance de celles-ci à l’intérieur de la tumeur. Ceci n’est pas le cas avec les MC-1 car elles sont à la fois anaérobies et magnétotactiques grâce à une chaîne de nanoparticules d’environ 70 nanomètres de diamètre, formant une sorte de « nano-boussole » magnétique à

(8)

l’intérieur de chaque bactérie, nous permettant de la diriger. Cette thèse étudie les différentes options de controle magnétique, et présente un système de contrôle magnétique programmable qui tient en compte d’une utilisation future chez les sujets humains.

De plus, la même chaîne de nanoparticules responsable du magnétotactisme, nano-cristaux d’oxyde de fer, constitue un agent de contraste en imagerie par résonance magnétique (IRM), qui est la modalité la plus efficace pour imager certains cancers tel que le cancer colorectal. Les MC-1, de par la taille de leur magnétosomes, engendrent une perte de cohérence du signal de résonance magnétique ce qui se traduit par une perte du signal sur l’image. L’étude des mécanismes de contrastes d’échantillons de MTBs révélent que les images en contraste T2 (temps de relaxation transversal) sont beaucoup plus affectées que celles en contraste T1 (temps de relaxation longitudinal) par la présence des MC-1. La quantité minimale qui peut être détectée avec un système standard d’IRM de 1.5Tesla est de 2.2×107_{MC-1/mL utilisant un contraste T2.} Nous avons aussi réalisé des expériences in vivo sur des souris porteuses de tumeurs afin de démontrer la capacité de transporter les MC-1 jusqu’à la tumeur. Les bactéries se sont retrouvées en grande quantité dans la partie nécrotique des tumeurs suite à une injection intraveineuse. L'application d'un champ magnétique focalisé sur la cible, en utilisant un système magnétique conçu et construit spécialement pour ces expériences in vivo, a permis d’accumuler les bactéries dans la tumeur. Dans cette thèse nous constatons, en moyenne, deux fois le nombre de bactéries quand un champ magnétique a été appliqué en utilisant un système de bobines 3D par rapport au transport passif avec le flux sanguin en l'absence de tout champ magnétique. Une deuxième expérience basée sur la comparaison de tumeurs implantées chez le même animal dévoile 10 fois plus de bactéries présentes dans la zone ciblée par rapport à celle non-ciblées, ce qui peut suggérer que le champ magnétique peut également être utilisé pour éviter une zone spécifique dans le corps, augmentant ainsi la specificité du traitement. Nous avons utilisé deux différents montages magnétiques pour le guidage des bactéries. Le premier montage, constitué principalement d’un électroaimant alimenté par une source de tension, est adéquat pour les tumeurs superficielles et accessibles. Le deuxième montage, constitué d’un ensemble de bobines tridimensionnel, permet un ciblage dans des sites profonds et inaccessibles permettant ainsi d’étendre l’application de l’utilisation des MC-1 pour le ciblage à plusieurs types de cancers dans le futur.

(9)

LIST OF TABLES

Table 1.1. Summary characteristics of some MTB ... 19 Table 3.1. Summary of some examples of the magnetic particles and the gradient required for their trapping ... 44 Table 6.1. Description of the groups used in the in vivo experiments. ... 71 Table 6.2. Summary of the key data obtained after histology slide analysis for the control and the experimental tumours. ... 74 Table 6.3. Summary of the key data obtained after histology slide analysis for the tumours implanted on the same mouse. ... 77 Table 7.1. Bacterial concentration and T2-relaxation values for different MTB concentrations calculated from signal ratio measurements. ... 85 Table 8.1. Summary characteristics of some MTB and calculation of the required gradient

necessary to have a magnetophoretic speed equivalent to the MTB speed giving a spherical magneitte microparticle with the same size range as the MTB. ... 95

(15)

LIST OF FIGURES

Figure 1.1. Differences between normal vasculature and one found in tumours. While normal vessels are regular and supply a homogenous blood flow to surrounding tissues, tumour angiogenesis blood vessels have dead ends, ramifications and an irregular size. Four distinct regions are noticeable where the treatment is perceived differently. The first region is a hypoxic necrotic one, the second is a semi-necrotic region, the third is a stabilized microcirculation region, and the fourth is an advancing front. Adapted from [30] ... 8 Figure 1.2. A computer controlled swarm of Magnetotactic bacteria was able to build a tiny pyramid made of SU-5 material blocks as depicted in A [47]. In B a scheme of a hybrid system made from synthetic material and propelled by the conjoint action of several bacteria attached to its end [42]. Scheme of a bacterium showing its body and an attached helical structure called flagella responsible of the bacterial motion through its spinning. The bacterial motion was reproduced by a synthetic system made from soft magnetic material head and a helical tail as shown in D [36]. Motion of this artificial flagella is granted through application of an alternating magnetic field generating a torque on the structure’s magnetic head as depicted in E [36]. ... 10 Figure 1.3. Principles of the magnetic drug targeting method. An active anticancer agent is encapsulated with a magnetic material forming a magnetoliposomes. Following an intravenous injection, retention of the magnetic particles that contain the drugs is achieved through the application of a strong magnetic field generated by an external source [57]. .... 12 Figure 1.4. Successful MRI-based navigation, targeting, and controlled release of a cancer drug in a pre-defined lobe in the liver of a live rabbit [64]. ... 13 Figure 1.5. Anaerobe bacteria colonize the necrotic central regions of solid tumours following an intravenous injection. The bacteria can be engineered to produce toxins that cause the death of cancer cells once in a poorly vascularized zone [67]. ... 15 Figure 1.6. Histology slide showing the presence of Salmonella inside a solid tumour [74]. ... 17 Figure 1.7. Electron scan microscopy images of MC-1 magnetotactic bacteria showing the round shaped cell with its internal organelle magnetosomes assembled into a chain. The MC-1

(16)

cells possess also two bundle sets of flagella. The magnetosome size range between 30-80nm [44]. ... 18 Figure 2.1. Bright-ﬁeld transmission electron micrograph (TEM) of a negatively stained cell of MC-1. The cell contains a single chain of magnetosomes and possesses two bundles of ﬂagellae (not visible on the image). There are 10-15 magnetosomes per bacteria, having an approximate mean size of 50nm. ... 23 Figure 2.2. Reference system for the simplified model describing the dynamics of a magnetotactic bacteria. The magnetic torque results from the angle between the magnetosomes chain and the external field while the hydrodynamic torque is produced by the movement of the flagella on the cell body. The cell body is represented by a sphere. .... 25 Figure 2.3. Representation of MC-1 in a 15 Gauss and 600 Gauss magnetic field. When subjected to a strong magnetic field, the angle of pitch of the helical motion increases causing the longitudinal speed of the bacteria to decrease. It was possible to record the helical motion of MTB by increasing the shutter speed of the Sony camera used to transfer video recordings to a computer. ... 27 Figure 2.4. In the absence of a magnetic field, MC-1 travels in a circular motion that is characteristic of non-magnetic bacteria swimming near boundaries. The zero magnetic fields were obtained by activating two pairs of magnetic coils with opposing current. In such a context, since the magnetic torque caused by the interaction of the magnetosomes chain with the external field is null, the hydrodynamic torque can be estimated with more accuracy. ... 28 Figure 2.5. Representation of three different configurations for the orientation of the magnetosomes chain influenced by an external field. A and B are the only possible configurations where the magnetic torque adds to the hydrodynamic torque. ... 29 Figure 2.6. Artistic view of polarity reversal once a bacterium hits the edge of a droplet. Because of the constant movement of the flagella when the edge is reached, the magnetosomes chain can be orientated in the opposite direction to the magnetic field. Single domain spheres assembled in a chain change polarity when submitted to an opposing magnetic field stronger than its coercive field. ... 30

(17)

Figure 2.7. Magnetotactic bacteria swim by means of flagella. Helical motion is a consequence of the non-integer number of turns of the flagella spire. MTB follow the magnetic field direction as in (a), however, near boundaries and obstacles, their motion forms an angle with the magnetic field. ... 31 Figure 2.8. Polarity selection magnetic setup. A sample containing both north-seeking and

south-seeking bacteria will be transformed into a homogenous sample containing only one population using this magnetic setup. The setup is guided by four permanent magnets arranged so that two positive magnets face each other and two negative magnets also face each other North-seeking bacteria will be directed to the center and will stay there. South-seeking bacteria will be directed to the edge of the setup and reverse direction due to the high magnetic field near the magnet surface, and will head to the center where the field is null. Since polarity reversal is permanent, only one population will remain. ... 32 Figure 3.1. Changing the current value for one coil of a Maxwell pair causes the position of the zero fields to move from the center. However, the gradient linearity is no longer preserved. ... 37 Figure 3.2. The relationship between the current IH flowing in a Helmholtz pair and the displacement value of the Maxwell magnetic fields. ... 38 Figure 3.3. Automatic navigation of MTB is done using a computer controlling the power supplies through a GPIB interface. ... 39 Figure 3.4. a) On the plane magnetic field velocity vector generated from 2D Maxwell pairs. b) Magnetic field absolute value as generated by a two Maxwell coil pairs in the x and y-axis. The magnetic field lines are directed toward the center where the field intensity is zero. .... 41 Figure 3.5. This figure shows an MTB sample under the influence of the Maxwell magnetic field. The bacteria from the entire sample are directed to the center of the container. The magnetic field gradient was 0.5Gauss/mm. ... 41 Figure 3.6. MTB guidance demonstration. In (a) and (b) a gradient was applied using the Maxwell coils. From c to f, the target location was changed which imply the activation of the Helmholtz coils. ... 43

(18)

Figure 4.1. Magnetic field generated by two pairs of orthogonal magnetic coils, powered by opposite current as shown in (a). The in-plane magnetic field is shown in (b), while the magnetic field in the orthogonal direction is shown in (c). The magnetic field direction in (a) converges to the center of the coils, causing the bacteria to aggregate in this point. In the orthogonal direction, however, the magnetic fields direction diverges. ... 49 Figure 4.2. Microfluidic channel used for in vitro bacteria targeting. The channel have a diameter of 100µm. There is two entry points to the channel as well as two exits (not shown in the picture). The channel was filled with PBS (Phosphate Saline Buffer), prior to the deposition of the bacteria in the insertion points. The bacteria are inserted without any pressure to ensure that the ones reaching the target are under the action of the magnetic field. ... 50 Figure 4.3. MTB distribution in a gradient magnetic field. The MTB lie inside the 0.3 Gauss magnetic equipotential circle, with more than half of them inside the 0.1 Gauss equipotential. ... 51 Figure 4.4. Targeting test in microfluidic channel. In (a) simulation of the magnetic field for a target location at the first U-shape occurrence in the channel. In (b) the corresponding MTB accumulation for the same magnetic field applied in a real channel with the same geometry. (c) When the converging field is set at another location, the bacteria are not able to bypass obstacle caused by the channel shape and to get to the target. ... 53 Figure 4.5. MTB magnetic control taking into account the geometry of the micro-fluidic channels. ... 54 Figure 4.6. Optical microscope images sequence showing the MC-1 MTB traveling in a microchannel following a given pattern. ... 54 Figure 4.7. Unblanced gradient from the two pairs of coils changes the shape of the target zone, delimited by the 0.3 Gauss equipotential. Far from the target, the magnetic field specifying the direction of the MTB changes helping the bacteria to avoid obstacles. ... 55 Figure 4.8. By using higher gradient, smaller targets inside the main one, can be specified, causing the far field to change direction. This technique can helo the bacteria to avoid obstacles. ... 56

(19)

Figure 5.1. Magnetic gradient field used for MTB aggregation and swarm formation. Once the bacteria enter the area where the magnetic field is less than 0.3 Gauss, they begin to move in circle. ... 62 Figure 5.2. Bacteria aggregate motion control along a pre-planned path. The circle with continuous line is the actual position of the bacteria, while the dotted-line circle represent the target position or the bacteria in the move. Aggregate formation is assured by the application of a gradient magnetic field. Shutting down the power for the coil4 result in displacement of the bacteria swarm along a funnel path. Considering the distance between the two positions and the mean bacteria speed, the time required for application of this field is determined. Once the time is up, the gradient field configuration is reapplied with unbalanced currents that cause the bacteria to aggregate in this new position. ... 64 Figure 5.3. Applying a magnetic field generated by 5 coils at a time multiplexed in order to cover all possible permutation resulted in a mean bacteria displacement toward the center of the coils. The time required for the aggregate formation using this control method is higher than formation of an aggregate in a 2D surface, because the applied magnetic field does not provide a direct path to the target for each instant of time. Again, powering the coils with unbalanced currents, cause the aggregate zone to be shifted. ... 66 Figure 5.4. A demonstration of bacteria control in 3D by multiplexing the power supplies over time. Powering only one pair of coils in a Maxwell configuration restrains the bacteria motion to a plane as depicted in (1). Photos (2) to (8) shows 3D control of a bacteria aggregate using the control method described in the text. ... 67 Figure 6.1. Magnetic field simulation showing the magnetic field lines as generated by (a) a 3D magnetic coils system and (b) an electromagnet or a permanent magnet. The magnetic field in (a) is focused on the target, and thus the bacteria will be attracted to the center of the tumour. However, in (b), the bacteria are attracted to the electromagnet tip. ... 72 Figure 6.2. Immunohistochemistry coloration of a section of the spleen and the liver displaying the presence of MC-1. ... 73 Figure 6.3. Graphic bar showing (up) the number of bacteria found in tumours for the 4 groups and (down) the invaded surface where a density of more than 20 MC-1 cells per 100µm2 was found. ... 75

(20)

Figure 6.4. In vivo MTB targeting experiment. MDAMB-231 tumour cells were grown in the flanks of nu/nu CD1 female mice. After MTB intravenous injection, a magnetic targeting procedure using an electromagnet (group III.A) and a 3D coils system (group III.B) was applied. An immunohistology analysis of the tumours reveals the presence of MTB in the necrotic region. The MTB exhibit a particular color as well as a characteristic size and shape which allow their differentiation it from the xenograft cells. A color map of the four experimental tumours groups is shown. The color corresponds to the bacterial count inside a FOV of 100µm2. We notice a wider distribution of MTB in the group III.A and group III.B compared to the group II. In fact, in group III.A and group III.B a magnetic guidance was imposed while for the group II only the blood flow is responsible for the delivery of MTB to the tumour. ... 76 Figure 6.5. Graphic bar showing the number of MC-1 cells for the tumours implanted on the same mouse. Each pair of bars represents respectively the non-targeted and the targeted tumour for one mouse. This experiment has been conducted only for the group III.A and group III.B. A higher magnetic targeting specificity is obtained with the 3D magnetic coil systems rather than the electromagnet. ... 78 Figure 7.1. Robotic platform depicting the MRI bore in which lies 3D Steering Gradient Coils (SGC) aimed to propel emboli magnetic particles. Facing the MRI bore, the 3D Steering Magnetic Coils (SMC) are responsible for MTB guidance and delivery to the tumour. The MRI table moves the patient from the inside of the MRI to the SMC. The magnetosomes inside the MTB cause signal loss on MR-images that allows, with the help of a proper calibration curve, the evaluation of the percentage of bacteria that reach the target. ... 81 Figure 7.2. Magnetic field lines generated by a magnetosome chain superimposed on top of an image of a magnetotactic bacterium. In this simulation, we consider the presence of 11 magnetosomes with a mean diameter of 70nm. The distance between the magnetosomes is of 20nm. The saturation magnetazation for magntite (0.6 Tesla) is considered since at the MRI field of 1.5 Tesla, the magnetite chain is saturated. ... 83 Figure 7.3. Simulation of a local magnetic field perturbation for a uniformly distributed concentration of MTB. The distance between MTB is taken to be 25µm, which corresponds to a concentration of approximately 107 bacteria/ml. ... 84

(21)

Figure 7.4. T1 images of several concentrations of magnetotactic bacteria (MTB) samples. Samples 1 to 6 (as numbered in a) show increasing concentrations starting with a medium without MTB (sample 1). T1-weighted spin echo sequence with TE = 11 ms, three different TR = 450/550/700 ms, a slice thickness of 20mm, and a pixel spacing of 0.586mm. (b) Signal intensity as a function of bacterial concentration for a T1-weighted acquisition with three different Repetition Time (TR) values. Notice that the signal contrast between samples of different concentrations is not important. ... 86 Figure 7.5. T2-images of several concentrations of magnetotactic bacteria (MTB) samples. Samples 1 to 6 (as numbered in a) show increasing concentrations starting with a medium without MTB (sample 1). T2-weighted fast spin echo sequence with TR = 5096ms, three different TE = 96/125/135ms, a slice thickness of 20mm, and a pixel spacing of 0.293mm. (b) Signal intensity as a function of bacterial concentration for a T2-weighted acquisition with three different Echo Time (TE) values. Notice that the signal contrast between samples of different bacterial concentrations is important. ... 87 Figure 7.6. T2-relaxation curves for different MTB concentrations. ... 88 Figure 7.7. Experimental T2 relaxation curves for magnetic and non-magnetic MC-1. The signal variation of the non-magnetic MC-1 is very similar to the medium. While the T2 value for the magnetic MC-1 was estimated to be 203ms for a bacterial concentration of 5⋅107, it was

found to be 725ms for the non-magnetic MC-1. The PBS medium has an experimental T2 equal 1072ms. T2 was estimated by fitting the signal intensity data for different TE values to a monoexponential decay curve. ... 89 Figure 7.8. Experimental T2 relaxation curves for motile and non-motile MC-1. Both samples show similar relaxation curves. An identical bacterial concentration of 1⋅108 MC-1 per ml

was used for both samples. The T2 value for the motile MC-1 was 162ms while it was 148ms for the non-motile MC-1. T2 was estimated by fitting the signal intensity data for different TE values to a monoexponential decay curve. ... 90 Figure 8.1. Y-shaped microvascular channel used for navigation simulation of steering efficiency. D represents the channel diameter and L its length; we consider L = D × 100. Even if

magnetic particle steering involves many forces, we consider only the most important, which are the magnetic and fluidic forces. MTB are subject to magnetic torque, but their

(22)

motion is governed by flagella, which allow the bacteria to move at constant speed along the direction of the magnetic field. ... 94

(23)

LIST OF ACRONYMS AND ABREVIATIONS

ABF Artificial Bacterial Flagellum COBALT Combination Bacteriolytic Therapy CPP Cell Penetrating Peptide

FOV Field of View GRE Gradient Echo

HCC Hepatocellular Carcinoma MNP Magnetic Nanoparticles MDT Magnetic Drug Targeting

MMP Magnetotactic Multicellular Prokaryote MRI Magnetic Resonance Imaging

MRP Magnetic Resonance Propulsion MTB Magnetotactic bacteria

PBS Phosphate Saline Buffer PEI Percutaneous Ethanol Injection

RF Radiofrequency

SE Spin Echo

SGC Steering Gradient Coils SMC Steering Magnetic Coils

SPIO Superparamagnetic Iron Oxides TEM Transmission Electron Microscopy

(24)

INTRODUCTION

Over the past two decades we have seen significant improvements in progression-free survival and reductions in morbidity for a number of cancer treatments, driven by the advent of biologically targeted therapies and more effective treatments. Many of these newer agents (e.g. Bevacizumab, Trastuzumab) actually sensitize tumours to the mainstay of anticancer therapy, cytotoxic chemotherapy. The latter is limited by significant acute and cumulative toxicities, such as myelosuppression and cardiac toxicity. While metastatic cancers most often require systemic therapy, there are specific contexts in which more effective localized therapy would have major impact on both quality of life and survival, including rectal and colorectal cancers, hepatocellular cancer, liver metastases from colorectal cancer and glioblastoma. We propose to use MC-1 magnetotactic bacteria (MTB) as a drug delivery system for solid tumours. MC-1 MTB use flagella for propulsion, while their swimming direction can be controlled by computer using an external directional magnetic field. This field induces a torque on a chain of nanoparticles that are naturally synthesized in the cell during cultivation. Like a compass needle, the swimming direction of the MTB is influenced through magnetotaxis. Furthermore, the MC-1 MTB spherical cells have a diameter of approximately 2µm, ideal for navigation in the smallest blood vessels, using an average swimming speed exceeding 200µm·s-1. Although the final aim of this research is to treat cancer, the bacteria used are not drug-loaded yet and do not produce any therapeutic effect. We are developing the concept of guiding the bacteria to solid tumours, and we are confident that once the proof of concept is achieved, following research projects will transform the bacteria for a therapeutic effect to take place.

The research hypotheses were as follows:

Hypothesis 1: It is possible to control the motion of a MC-1 MTB swarm in a tridimenstional space under computer control.

Hypothesis 2: MC-1 MTB cause a significant magnetic resonance signal distortion making their targeting possible using clinical MRI systems.

(25)

We will address three main objectives in this thesis: Objective 1: development of a programmable MC-1 MTB magnetic control system. This system has to be designed for a human patient even though there are no planned human trials. Objective 2: characterization of the contrast caused by the MC-1 MTB on magnetic resonance images. The magnetic control system combined with the magnetic resonance imaging will form the future platform for MTB guidance. Objective 3: assess and quantify the MC-1 distribution in tumour xenografts and evaluate the magnetic guidance advantage over passive targeting.

Chapter 2 of this thesis studies the hydrodynamics of the MC-1 magnetotactic bacteria in high magnetic fields. It elaborates the influence of the magnetic torque on the overall motion of the bacteria and the consequences following application of a high magnetic field on a bacterial sample such the one found in MRI systems. Increasing the magnetic field leads to a decrease in speed of the MTB. During these observations, the magnetic field up to 1.27Tesla was used, which is quite similar to the one found in most MRI systems (1.5Tesla). Most importantly, MTB show a change in polarity when the field is released; for example, if a sample of 100% north-seeking bacteria is submitted to a 1 Tesla field, the MTB will be divided into 50% north north-seeking and 50% south seeking bacteria once the field is released. This observation greatly impacts the targeting method as we have suggested at the beginning of the project to use the MRI image as a feedback to correct the magnetic field used to control the bacteria. In fact, we can conclude that doing this will change the bacteria’s behavior, which will have a negative impact on controllability during in vivo drug delivery applications. A mathematical model describing the interaction between the magnetosome chain and the magnetic field and how it affects motion is presented along with a an explanation on what causes the bacteria to change polarity when they hit an obstacle.

Chapter 3 covers the different MTB control techniques that are suitable for in vivo drug delivery applications. After several iterations, designs and much experimentation, we found that the coil structure providing the best results for MTB navigation was a mix of 3-axis Helmholtz coils and 3-axis Maxwell coils. The Maxwell configuration traps the MTB that follow the magnetic field lines in the center of the coils. Controlling the motion of the bacteria then becomes possible using the computer-controlled Helmholtz coils. The current flowing in the Helmholtz coils and the position of the zero fields generated from the Maxwell coils are linearly related.

(26)

In chapter 4, we characterize the swimming of MTB in vitro by mimicking different situations they could encounter in an in vivo environment. The goal was to be able to predict the MTB behavior in a complex microvasculature and to choose the correct magnetic field algorithm to optimize targeting. In fact, as the MTB follow the magnetic field, they could easily get stuck on obstacles, especially in very complex capillary networks such as the tumour angiogenesis. Consequently, changing the direction of the field is necessary to guide the bacteria. The problem is a lack of visual information on the route and obstacles that the bacteria face.

Since we lack this information, we must achieve blind control knowing only the starting position of the bacteria (the injection site) and their final destination (the center of the tumour). As a simple example to help understand the problem and the way we propose to solve it, we used a microchannel design consisting of multiple U-channel shapes. The bacteria were injected into two tanks connected to the network through a small capillary. The magnetic field was set to a target location. We measured the number of bacteria that reached the target as time passed. We show in this paper that a magnetic field algorithm is very important in order to navigate the bacteria accurately and that it can be generalized to complex geometries. Taking into account the geometry of the channels, the bacteria’s speeds and behavior for a given magnetic field, we can optimize targeting with the appropriate magnetic field. Modeling and simulation can play an important role in this case. In fact, using the typical architecture of angiogenesis found in scientific literature in addition to the characterization of bacteria when submitted to different environmental parameters, we can predict and engineer a magnetic field algorithm that would optimize targeting. The direction, intensity and duration of the magnetic field are the three parameters to consider for MTB guidance. Changing the direction of the magnetic field allows the MTB to bypass obstacles while modulation of its intensity will give more or less freedom to the MTB to deviate depending on oxygen or other physical entities.

The paper presented in chapter 5 extends the usage of the magnetic control method presented in chapter 5 to a three-dimensional space and presents the required time multiplexing of coil powering in order to achieve bacterial swarm control in 3D. The notion of a magnetic monopole does not exist in nature and it is impossible to reproduce in laboratory. In fact, the magnetic field that converges in a two-dimensional space exits from the third dimension. We propose in this paper a method that allows us to retain a swarm of bacteria in a target location.

(27)

Chapter 6 presents the in vivo results of MTB delivered to a tumour xenograft implanted on a mouse. Two different magnetic setups were used for the delivery of MTB to the tumour following their intravenous injection. The first setup consists of an electromagnet placed as near as possible to the tumour, while the second involves the use of the tridimensional-coil system described in chapters 4 and 5. Since MTB have a life span of 30-45 minutes at body temperature, the duration of the experiment was set to 30 minutes during which the magnetic field was applied to the tumour. Following the experiment, the animal was euthanized and the tumour extracted for analysis. Histochemical staining allowed identification of the bacteria on the histology slide.

(28)

CHAPTER 1 LITERATURE REVIEW

This literature review is comprised of four parts. The first part covers the main cancer therapy methods as well as the major physiological barriers that drugs encounter on their way to targeted tissues. The goal of the first part of this review is to bring to the reader’s attention the importance of developing a tool that deliver the active drug directly to tumours. The second part discusses methods of delivery, namely the magnetic drug targeting approach and the Magnetic Resonance Propulsion (MRP) method that was first proposed and developed by the NanoRobotics laboratory of the Ecole Polytechnique de Montreal. The third part covers the use of bacteria in cancer treatment and focuses on their application clinical applications. Finally, the last part of the review presents the MC-1 magnetotactic bacteria (MTB) and discusses particularities of the MTB that are accessible in the literature.

1.1 Cancer Therapy Approach

Cancer, as well as the therapy undergone following its diagnosis, is classified depending on where it develops in the body. For example, patients with colorectal cancer1 will undergo integrated surgery and chemotherapy whereas in the case of patients with rectal cancer, additional radiation therapy is used. On the other hand, hepatocellular carcinoma2 (HCC) requires resection, liver transplantation, and percutaneous ablation (percutaneous ethanol injection (PEI) and radiofrequency (RF)) that can only be used on 30-40% of the patients. Meanwhile, chemoembolization is the only alternative therapeutic strategy capable of improving survival for the majority of patients. Chemoembolization combines arterial administration of drugs with a form of arterial occlusion in the liver by an embolizing agent. Systemic therapies, which affect the entire body, are used for the treatment of almost all type of cancer.

1_{Colorectal cancer is the third most prevalent malignancy worldwide and the second leading cause of cancer death in}

Canada based on Canadian Cancer Statistics 2009.

(29)

1.1.1 Systemic therapies

The most common systemic therapies are radiotherapy, chemotherapy, immunotherapy, and gene therapy. Radiotherapy involves irradiation of the tumour by ionization radiation that damages the DNA of cancerous cells. It is used for patients who are not permissible for surgery, or as a complementary treatment. The source of radiation can be external or internal, which mainly depends on the affected organ. Image-guided radiotherapy accounts for the body motion, such as respiration and internal organs motion, through the use of advanced imaging modalities. The efficiency of radiotherapy is however hindered by the hypoxic nature of the center of solid tumours that poses a limitation since the cells in this region are 2 to 3 times more resistant to radiation than normal cells [1, 2]. This region of the tumour also decreases the efficacy of Chemotherapy, which relies on systemic blood circulation to transport drugs. Chemotherapy acts by killing rapidly dividing cells such cancerous ones, and as a matter of fact, quiescent cells, which are far from the vasculature, are hardly affected [3]. In order to reduce its side effects, the active agent is encapsulated inside particles that attach to certain proteins present on cancerous cells and are subsequently actively targeted at tumours [4]. Active targeting is not to be confused with direct delivery, which encompasses methods that physically transport the active agent to the tumour. Immunotherapy enhances the way the immune system deals with cancerous cells by inducing a response against the antigen they express [5]. It consists of injecting lymphocyte cells that are “trained” to recognize and destroy specific cancer cells. Gene therapy uses a vector, usually a virus, to transport and inject healthy human genes inside cancerous cells. The short circulation time of vectors as well as their toxicity, immune system responses, gene control and targeting issues all prevent successful medical application of this technique [6]. The success of the techniques mentioned above depends on the efficacy of the active agent in killing cancerous cells as well as the quantity that reaches the tumour. The injected dose is limited by the tolerance of the body since systemic therapies affect cancerous cells as well as healthy cells [7, 8]. Moreover, tumours induce many physiological barriers that hinder the drugs from reaching the target.

1.1.2 Physiological barriers

The systemic therapies described in the previous section rely on transporting particles, molecules or cells to the tumour. For the therapy to be successful, the active agent must reach the tumour in

(30)

sufficient quantity. First, the active agent has to reach the main blood vessels that lead to the tumour; second, it must extravagate in the interstitial space of the tumour, and then migrate so as to be in direct contact with the cancerous cells [9-15]. The heterogeneous blood flow of the tumour and the high interstitial pressure are the major obstacles that drugs need to surmount to reach the target [16-20].

1.1.2.1 Tumour blood supply and transvascular transportation in tumours

The rapid growth of the tumour generates a lack of oxygen and nutrients and triggers the creation of new blood vessels, which is known as angiogenesis [21, 22]. Blood vessels thus created are different from healthy ones, and are characterised by leakiness, ill formation, and inhomogeneities [23-25]. In fact, large pores exist in tumour blood vessels [26]. The pore size depends on the tumour itself as well as nearby organs [14, 15, 19, 20, 24], it can range from a nanometer, as is the case with human glioblastoma (HGL21) transplanted in the cranial windows of a mouse [27], to 2000nm for the mammary adenocarcinoma of a mouse (MCalV) transplanted into the dorsal skinfold chamber [28]. Moreover, tumour angiogenesis blood vessels have dead ends, ramifications and an irregular size [11, 12, 16-18]. Heterogeneities characterise transvascular transport in tumours, as shown experimentally in [15], where fluorescent liposomes that were injected in tumour-bearing mice accumulated in some regions of the tumours but not in others. These characteristic particularities of tumour blood supply dictate the way drug molecules reach the tumour cells. In addition, systemic therapies have different therapeutic effects depending on the perfusion of the tumour. In fact, there are four distinct regions where the treatment is perceived differently as depicted by Figure 1.1. The first region is a hypoxic necrotic

one, the second is a semi-necrotic region, the third is a stabilized microcirculation region, and the fourth is an advancing front [10, 29].

(31)

Figure 1.1. Differences between normal vasculature and one found in tumours. While normal vessels are regular and supply a homogenous blood flow to surrounding tissues, tumour angiogenesis blood vessels have dead ends, ramifications and an irregular size. Four distinct regions are noticeable where the treatment is perceived differently. The first region is a hypoxic necrotic one, the second is a semi-necrotic region, the third is a stabilized microcirculation region, and the fourth is an advancing front. Adapted from [30]

1.1.2.2 Interstitial transport in tumours

Molecules or particles that successfully cross the vascular wall will encounter a second major obstacle before reaching cancerous cells. Transport through interstitial space poses a real challenge for passive particles because of the high interstitial pressure inside the tumour and the lack of functional lymphatic vessels [31]. However, when the interstitial space of tumours is evaluated, it is found that it is much larger than those found in normal tissue [32, 33]. While this observation might suggest that transporting molecules and particles should be easier in tumour tissue than in normal tissue, experimental data shows the opposite. In fact, the interstitial hydrostatic pressure is high in the center of the tumour and low in the periphery. Therefore, interstitial fluid motion is expected to move from the center of the tumour to the periphery. Various studies show that 1-14% of plasma entering the tumour leaves through the periphery

!"#$%&'(%)*+&%,+#-''

.+$"#'(%)*+&%,+#-'

/01"23*'4"5-'

6+3-)*-5,'4"5-'

(32)

[34]. The outward fluid velocity resulting from this leakage is estimated to be 0.1-0.2 µm/s at the periphery of 1 cm of tissue-isolated tumour [9]. Thus, particles or molecules must counter this fluid motion in order to reach the interior of the tumour.

Diffusion and convection govern passive transportation of molecules in the interstitial space. Furthermore, because of micro and macroscopic heterogeneities in tumours, the magnitude of these two parameters varies inside the same tumour and over time as well as from one tumour to another [35]. Active transport, such as the use of micro-engineered systems, magnetic particles steered by magnetic fields, or bacteria may enhance the transportation of drugs in transvascular and interstitial spaces, as will be discussed in the next section.

1.2 Direct drug delivery to tumours

Transportation of untethered microcarriers bearing therapeutic agents through the vascular network of a human body to a target requires an appropriate propulsion and steering system. The development of such propulsion and steering system can be a real technical challenge especially when the dimensions of these microcarriers must be reduced to a diameter of approximately 2µm to allow them to operate efficiently in the narrowest blood vessels of the microvasculature. An entirely synthetic approach that does not rely on an external source of induced propulsion force is beyond the limit of current technology. In fact, approaches that do rely on an external source for propulsion are still facing many technological challenges when designed to operate in the human microvascular networks. The major limitation consists of difficulties in miniaturising the power source that allows efficient autonomy of the microcarrier.

1.2.1 Synthetic systems

Flagellae and cilia, nanomotor-like mechanisms by which most microorganisms move, have inspired researchers in the design of modern engineering tools. Artificial Bacterial Flagellum (ABF) made from a helical tail, resembling a natural flagellum, and a soft-magnetic material head was fabricated in [36], based on a principle that was theoretically described in [37]. Under the action of an alternating magnetic field, the ABF was successful in reproducing the bacterial motion using a synthetic system [38, 39] as depicted by Figure 1.2. However, profiting from the

existing flagella of microorganisms such as bacteria appeared more advantageous [40-44]. Successful on/off motion control of the bacteria by adjusting the chemical composition of their

(33)

medium was demonstrated in [45]. In addition, bacteria have shown to possess sufficient thrust force to move micro-objects from one location to another [46], and to build more complex structures under the conjoint action of a swarm of microorganisms [47].

Figure 1.2. A computer controlled swarm of Magnetotactic bacteria was able to build a tiny pyramid made of SU-5 material blocks as depicted in A [47]. In B a scheme of a hybrid system made from synthetic material and propelled by the conjoint action of several bacteria attached to its end [42]. Scheme of a bacterium showing its body and an attached helical structure called flagella responsible of the bacterial motion through its spinning. The bacterial motion was reproduced by a synthetic system made from soft magnetic material head and a helical tail as shown in D [36]. Motion of this artificial flagella is granted through application of an alternating magnetic field generating a torque on the structure’s magnetic head as depicted in E [36].

1.2.2 Magnetic drug targeting

Magnetic Drug Targeting (MDT) is a technique that magnetically concentrates drugs at the tumour site in order to reduce secondary toxicity effects that follow common therapies in cancer treatment [48-52]. Currently, MDT consists of loading drugs into magnetic particles that carry this complex in the vicinity of a tumour using an external applied magnetic gradient as described by the Figure 1.3. Moreover, lack of navigational control when combined with complex

A C

B

D

(34)

microvascular networks near a tumour contributes further in lowering the efficacy of such targeting.

Present implementations for magnetic targeting of tumours rely on an external permanent or an electro-magnet located near and above the tumour. In this method, a catheter is typically used to release the magnetic nanoparticles as close as possible to the target. However, due to higher field intensity towards the external magnet [53], targeting is mostly restricted to superficial tumours near the skin. Because of higher magnetic gradient intensity towards the external magnet, targeting efficacy is higher for tumours located near the magnet and decreases substantially when the tumour is located deeper in the body.

Therefore, as the targets get deeper in the body, significant reduction of targeting efficacy is anticipated with this technique. Furthermore, this approach relies on trapping the particles without any further navigation or trajectory control over pre-planned pathways towards the tumour; thus, the distance between the nanoparticles releasing site and the tumour significantly affects targeting effectiveness. The dimensions and the technology of the available catheterization will always limit the efficacy of this technique to reach a desired site in the complex microvasculature. Deep organ targeting has been improved through the use of magnet tipped catheter, magnetic needles, wires or stents [48, 54-56], but still, the non-linear geometry of the induced field and the resulting distribution of the particles remain uncontrolled.

When tumours near the surface are considered, such as head, neck and skin carcinomas, experimental data showed complete remission of the tumour after one treatment using magnetic particles. This significant achievement lead to better efficacy with only one-fifth the medication that would otherwise conventionally be needed. Combining MDT with hyperthermia is yet another approach to enhance efficacy of a well-established tumour treatment. Involvement of big companies such Siemens provide a positive outlook for the clinical future of this technique.

(35)

Figure 1.3. Principles of the magnetic drug targeting method. An active anticancer agent is encapsulated with a magnetic material forming a magnetoliposomes. Following an intravenous injection, retention of the magnetic particles that contain the drugs is achieved through the application of a strong magnetic field generated by an external source [57].

1.2.3 Magnetic Resonance Propulsion (MRP)

In order to improve targeting efficacy for tumours located deeper in the body, a new method has been proposed. This method relies on an induced propulsion force on magnetic materials (with sufficient magnetization saturation) generated by the three orthogonal coils used for image encoding in conventional clinical Magnetic Resonance Imaging (MRI) systems. [58-64] Using the tracking information provided by the imaging modality of the MRI, a closed-loop computer controlled displacement of a ferromagnetic bead along a pre-planned pathway in the vasculature of a living animal was demonstrated. The experiment was conducted in the carotid artery of a living swine and the results illustrated that this technology was able to adapt the navigation of magnetic microparticles through the microvasculature for future therapeutic applications beyond the limits of modern catheterization. This experiment also ensured that using MRI as an imaging modality would not only be advantageous for the purpose of tracking of the magnetic

(36)

microparticles, but also for real-time monitoring of therapeutic efficiency during such interventions [61].

Figure 1.4. Successful MRI-based navigation, targeting, and controlled release of a cancer drug in a pre-defined lobe in the liver of a live rabbit [64].

Inducing sufficient force on nanoparticles would require magnetic gradients with amplitudes that are beyond today’s technological advances when operating at the human scale. This is due in great part to the very small volume of the magnetic materials involved. Even if such high amplitude gradient field was achievable, there are physiological limitations to the gradient field amplitude that have previously been pointed out. One possible solution when operating in the microvasculature is to encapsulate such magnetic nanoparticles with therapeutic agents within polymeric microcarriers [64] as shown in Figure 1.4. With this approach, a higher effective volume

of magnetic material can be achieved while retaining the advantages offered by the properties of magnetic nanoparticles for MR-target applications. However, tumour lesions are typically accessible by transiting through anarchic arteriocapillar networks stimulated by tumour angiogenesis. These capillaries are as small as 4-5µm in diameter. As such, this environment restricts the overall maximum diameter of each magnetic microcarrier for efficient navigation to approximately 2µm. Although at such a scale, the blood flow would typically be used for propulsion, sufficient magnetic gradient must still be generated to steer such microcarriers efficiently, especially at vessel bifurcations. Preliminary studies showed that additional gradient steering coils could provide maximum gradients sufficient for larger microcarriers used for target chemo-embolization but may prove to be insufficient for enhanced targeting inside a tumour. One approach proposed by our group is to eliminate the need for such gradient coils when operating in the microvasculature and to replace it by the propulsion force provided by flagellated

(37)

bacteria. Since steering control is also required for target interventions, special types of cells referred to as Magnetotactic Bacteria (MTB) are considered by our group for such applications.

1.3 Bacterial cancer therapy

Bacteria that are present in the human body outnumber human cells by a factor of 10 to 1 [65]. Relying on complex mechanisms for their motion, sensory apparatus and communication, they play important roles in many aspects of the body’s operation. Ever since a patient accidentally infected by bacteria recovered from cancer in 1891 [66], many researchers have focused on the use of bacteria as a potential treatment for cancer [6, 67-74]. The phenomenon by which the bacteria helped the patient to recover from cancer remains unknown, though many researchers believe that their ability to proliferate in the center of a solid tumour, and their production of toxins figure as key factors. Among the tremendous number of bacteria available in nature, the anaerobic ones (i.e. bacteria that live in the absence of oxygen) have been the primary consideration for most researchers that have tried to reproduce the accidental success witnessed by Dr. Coley [66, 73]. Figure 1.5 shows an example of an application for the bacterial cancer

therapy.

1.3.1 Principle of bacterial therapy

Bacteria are believed to be more specific when targeting hypoxic regions of the body that are characteristic of solid tumours [68]. The commonly used bacteria belong to the Clostridia,

Salmonella and Bifidobacterium strains [75]. They can be genetically modified to express a

certain gene that can have an antitumour effect or supress certain genes to make the bacteria better tolerated by the body. The therapeutic effects come from the bacteria thriving from consuming necrotic cancerous tissue, or through products of bacterial activity such as endotoxins, which can be used for tumour destruction. Bacteria can also be used as delivery agents for anticancer drugs, and as vectors for gene therapy.

(38)

Figure 1.5. Anaerobe bacteria colonize the necrotic central regions of solid tumours following an intravenous injection. The bacteria can be engineered to produce toxins that cause the death of cancer cells once in a poorly vascularized zone [67].

1.3.2 Tumour-colonizing bacteria

It was found through animal model experiments that, following an intravenous injection, anaerobic bacteria found an appropriate environment for their proliferation in the hypoxic region of solid tumours as shown in an histology slide in Figure 1.6. The presence of pathogenic

anaerobic clostridia species resulted even in tumour regression, but caused a severe toxicity, which resulted in either death or illness of the host animal [76, 77]. Non-pathogenic strains of

Clostridium were also able to colonize the hypoxic zone of solid tumours, yet they did not cause

tumour regression as the pathogenic strain did [78]. An attenuated strain, the C. novyi-NT, which was obtained after deleting gene coding for lethal toxin, caused tumour regression, but the